The Pulse
Satlyt Tests Orbital Software for Shared Computing
Satlyt tested software aboard a spacecraft to plan how computing tasks could be assigned across an orbital network. The test modeled four computers but used only one physical computer in orbit, leaving transfers between spacecraft for futur
AI.info Team · Written with AI · fact-checked ·
“It’s a building block toward satellites sharing computing work.”
Rama Afullo, founder and CEO of Satlyt
Satlyt tested software aboard a spacecraft earlier in October that plans how computing tasks could be distributed across an orbital network. The experiment, conducted with the University of Houston as part of a NASA-funded research project, showed the software could produce task assignments using computing capacity and communication availability. It did not demonstrate computers on separate spacecraft sharing work: only one computer physically ran the experiment in orbit.
One onboard computer modeled a four-computer network
For the test, Satlyt set up a simulated network of four computers, four users and eight services. Its software assessed how to assign and run tasks across the modeled computers, then returned the results to Earth. The actual onboard hardware evaluated the simulated network; the other computers were not physical nodes exchanging data in orbit.
That distinction marks the experiment’s limit. Satlyt showed its software could generate plans on spacecraft hardware, but it has not yet demonstrated how multiple physical computers transfer information between one another in practice. Afullo said one approach performed better than the others, though it also required more computing power.
NASA funded the research, not a Satlyt deployment
The project’s NASA funding does not mean the agency has selected Satlyt for a mission or chosen to deploy its technology. TechCrunch reported that Satlyt’s software is already on three satellites in orbit, but the test described in the story focused on running its planning software on one onboard computer.
Satlyt’s wider aim is to connect spacecraft so they can share computing assignments. Operators would set permissions, priorities and operating limits; the software would make decisions within those boundaries. Processing data near where a satellite collects it could reduce reliance on downlinks to Earth, which can be constrained by limited communication windows.
Afullo founded Satlyt in 2024 to develop software that lets operators run AI and other applications aboard spacecraft. One example he gave is a satellite detecting a wildfire and sending relevant information sooner, rather than waiting to transmit all collected data for ground-based analysis. He said the experiment tested one planning component of that broader capability.
A timing error and several engineering hurdles remain
The test also exposed an issue in how Satlyt’s software referenced delays: it recorded arrival time rather than elapsed delay. Those are different measurements, and Afullo said the company still needs to test the software across varied workloads, hardware and connectivity conditions.
Before computing work can be demonstrated across separate spacecraft, Satlyt must also address communication links, operator permissions, secure transfers, resource scheduling and recovery when a connection becomes unavailable. Afullo said the next steps are broader validation and integration, followed by demonstrations that share computing work across spacecraft. The October test returned its outputs to Earth, but it did not yet show that cross-spacecraft exchange.